1 / 24

Image Compression

Image Compression. CSC 1040. Image Compression. Goal: reduce redundancy Encode the same information using fewer bits Originally developed for fax transmission Send high quality documents in short calls Image compression can lose information by approximating colors:

mya
Download Presentation

Image Compression

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Image Compression CSC 1040

  2. Image Compression • Goal: reduce redundancy • Encode the same information using fewer bits • Originally developed for fax transmission • Send high quality documents in short calls • Image compression can lose information by approximating colors: • Lossy: can’t reconstruct original image, but looks the same • Lossless: image can reconstructed exactly CSC 1040 - Computing with Images

  3. Human Vision • Human eye tends to “blend” nearby colors • Visual acuity varies markedly across features • Discontinuities easily seen, absolutes less crucial • High frame rates produce apparent motion • Smooth motion requires about 24 frames/sec CSC 1040 - Computing with Images

  4. Do these colors look the same as ... CSC 1040 - Computing with Images

  5. ... as these? CSC 1040 - Computing with Images

  6. Not quite CSC 1040 - Computing with Images

  7. Two main techniques • Run-length encoding • Group areas of similar pixels together and encode their color only once • Reduces size as if image had fewer pixels • Palette selection • Encode only the colors actually used in the picture (may be much fewer than total number of possible colors) • Reduces size by using fewer bits per pixel CSC 1040 - Computing with Images

  8. Run length encoding • Opportunity: • Large regions of a single color are common • Approach: • Record # of consecutive pixels for each color • An example of lossless encoding 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Uncompressed 000000000000000000000111111111111100000000000001111111111111111111111 RLE Row 1, 21:0,13:1;13:0;22:1 LZW, etc. use algorithms in addition to RLE 01010101010101010101 CSC 1040 - Computing with Images

  9. JPEG • Joint Photographic Experts Group • Opportunity: • Eye sees sharp lines better than subtle shading • Approach: • Retain detail only for the most important parts • Accomplished with Discrete Cosine Transform • Allows user-selectable fidelity • Results:Typical compression 20:1 CSC 1040 - Computing with Images

  10. JPEG • Joint Photographic Experts Group • Preferred format for scanned photographic images for use over the internet or Web. Not meant for printing. • Not good for images with a lot of solid color, vector drawings, type, or line art or images with “Web-safe” colors. • JPEG compression is lossy! Save and archive the original before converting to JPEG. CSC 1040 - Computing with Images

  11. In Photoshop, when you Save as… a JPEG file, you can choose the level of compression and, therefore, the size and quality of the file. CSC 1040 - Computing with Images

  12. Palette selection • Opportunity: • No picture uses all 16 million colors • Human eye does not see small differences • Approach: • Select a palette of 256 colors • Indicate which palette entry to use for each pixel • Look up each color in the palette CSC 1040 - Computing with Images

  13. Size of images Same as grayscale Half of grayscale CSC 1040 - Computing with Images

  14. GIF • Graphics Interchange Format • Palette selection, then lossless compression • Opportunity: • Common colors are sent more often • Approach: • Use fewer bits to represent common colors. Example: • 1 Blue 75% 75x1= 75 75x2=150 • 01 White 20% 20x2= 40 20x2= 40 • 001 Red 5% 5x3= 15 5x2= 10 • 130 200 CSC 1040 - Computing with Images

  15. GIF • Graphics Interchange Format • Industry standard graphic format for on-screen viewing through the Internet and Web. • Not meant to be used for printing. • The best format for all images except scanned photographic images (use JPEG for these). • GIF supports lossless LZW compression. CSC 1040 - Computing with Images

  16. “Lossy” compression types: • JPEG • Genuine Fractals • Mr. SID 9:1 “averaging” Original Compressed CSC 1040 - Computing with Images

  17. Tools details for this option Photoshop Save for Web settings CSC 1040 - Computing with Images

  18. JPEG cautions: • Images with hard edges, high contrasts, angular areas, and text suffer from JPEG compression. • Scanned “natural” photographs do not lose much, especially at High or Maximum quality. • Only save finished images as JPEGs, every time you open and save again, even if you don’t edit, you lose quality. • Always keep the original non-JPEG version (the native .psd format). • So why use JPEG? • It is the best format for photographic images on the Web. • Its compression ability is very great. CSC 1040 - Computing with Images

  19. GIF cautions: • Restricted number of colors – not suitable for photographs • Very good compression for line drawings (uniform color areas) • Only save finished images as GIFs • Always keep the original non-GIF version (the native .psd format). • So why use GIF? • It is a good format for images that have few colors and is widely available • Can do lossless compression • Its compression ability is great, especially for images with few colors • Can also be used for animation CSC 1040 - Computing with Images

  20. printing web Adobe Photoshop Photoshop can save files in many file formats: .psd — Photoshop Document (“native” format) .pdf — Photoshop Portable Document Format .eps — Photoshop Encapsulated PostScript .tiff — Tagged Image File Format .gif — CompuserveGraphic Interchange Format .jpg — JPEG, Joint Photographic Experts Group .bmp — Windows Bitmap .png — Portable Network Graphic etc., etc. A summary of Graphics file formats and uses can be found in: http://en.wikipedia.org/wiki/Graphics_file_format_summary CSC 1040 - Computing with Images

  21. EPS • Encapsulated PostScript • Preferred file format for importing into page layout programs such as QuarkXPress, PageMaker, InDesign, XPress, etc. for subsequent printing. • An object-oriented format. • Will only print to a postscript printer. • Uses lossy JPEG compression. • Only save your file as EPS if you need to import it into a page layout program. CSC 1040 - Computing with Images

  22. TIFF • Tagged Image File Format • Widely used cross platform file format also designed for printing. • A bitmap image format. • TIFF supports lossless LZW compression which also makes it a good archive format for Photoshop documents. CSC 1040 - Computing with Images

  23. PDF • Portable Document Format • Developed to transfer and read documents without having to print them—the “paperless office.” • Cross platform format that can be read with the free download Adobe Acrobat Reader. • Can represent both vector and bitmap graphics. • Can also contain electronic document search and navigation features as well as hypertext links. • Can be created from almost any application, but the user cannot edit or modify the file except with Adobe Acrobat (or other software). • Document formatting, fonts, colors, etc. are maintained and appear identical across platforms. • Excellent in the “prepress” process — can be sent to the printer, but can also be placed in other documents. CSC 1040 - Computing with Images

  24. Some of the slides in this presentation were adapted from: http://www.entomology.umn.edu/museum/links/coursefiles/ENT 5051/PowerPoint presentations/Lec 7-File formats.ppt and http://www.umiacs.umd.edu/~jimmylin/LBSC690-2004-Fall/Week10.ppt CSC 1040 - Computing with Images

More Related